Seasonal and Interannual Correlations Between Right-Whale Distribution and Calving Success and Chlorophyll Concentrations in the Gulf of Maine, USA

Vol. 394: 289–302, 2009 MARINE ECOLOGY PROGRESS SERIES Published November 18 doi: 10.3354/meps08267 Mar Ecol Prog Ser

OPEN ACCESS

Seasonal and interannual correlations between right-whale distribution and calving success and chlorophyll concentrations in the Gulf of Maine, USA

Brittan L. Hlista1, Heidi M. Sosik1,*, Linda V. Martin Traykovski1, Robert D. Kenney2, Michael J. Moore1

1Biology Department, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543, USA 2Graduate School of Oceanography, University of Rhode Island, Narragansett, Rhode Island 02882-1197, USA

ABSTRACT: The North Atlantic right whale Eubalaena glacialis is one of the most endangered species of large whales. Although human-caused mortality is the primary factor contributing to poor recovery of E. glacialis, variability in reproductive success may also play a role. The present study evaluates the idea that seasonal distributions and reproductive success in E. glacialis are linked to food availability or related environmental conditions that can be assessed by treating satellite- derived sea-surface chlorophyll (chl) concentration as a proxy. Sea-surface chl time series in the major high-use feeding habitats and whale-sightings data were compared. Whale transition between habitats reflects a pattern in the seasonal distribution of peak concentration in satellite-derived chl. A regionally and seasonally weighted chl index was calculated to reflect aspects of average potential food condition. We found a significant correlation between the number of whale calves born and the weighted chl averaged over the prior 2 yr. These findings are consistent with the view that food availability during and just before the gestation period may be a critical factor regulating reproductive success, with low food years contributing to delays in conception. Longer time series are necessary to examine the predictive relationship between weighted chl concentration and calf production. Although ecological interactions and whale reproductive biology are certainly more complex than can be encompassed by emphasizing only food availability, analysis of satellite-derived surface chl concentrations provides a practical means to monitor a level of ecosystem variability that affects right-whale distributions and productivity.

KEY WORDS: Right whale · Eubalaena glacialis · Calving · Seasonal patterns · Ocean color · SeaWiFS · Chlorophyll · Remote sensing

Resale or republication not permitted without written consent of the publisher

INTRODUCTION 5.3% in 1980 to –2.4% in 1994 (Caswell et al. 1999). Overall, the E. glacialis population growth rate is As a consequence of the whaling industry, the North extremely low (<1% since 1992; Fujiwara & Caswell Atlantic right whale Eubalaena glacialis population 2001). The 2 major anthropogenic sources of right- was reduced to very low levels by the end of the 19th whale mortality include vessel strikes and entangle- century and its current population is estimated at ment with fixed fishing gear (Knowlton & Kraus 2001). <400 ind. (IWC 2001, Kraus et al. 2005). Despite inter- Although anthropogenic mortality compromises the national protection from commercial whaling since recovery of this population, strong evidence suggests 1935, the population is showing slow signs of recovery, that a variable reproductive rate (Knowlton et al. 1994, at best. Demographic estimates have indicated a grad- Kraus et al. 2001) and a calving interval that fluctuates ual decline in the population growth rate from about from 3 to 6 yr (Kraus et al. 2007) also contribute to the

*Corresponding author. Email: [email protected] © Inter-Research 2009 · www.int-res.com 290 Mar Ecol Prog Ser 394: 289–302, 2009

reduced population growth rate. The factors that instance, an anomalous winter bloom across the Gulf of could contribute to decreases in reproductive rate Maine in late February 1999 may have allowed an include environmental contaminants, body condition/ extra generation of C. finmarchicus to develop, prece- nutritional stress, loss of genetic variability, infectious ding the usual March-April spring bloom, thus leading disease, and marine biotoxins (Reeves et al. 2001). to a substantial accumulation of the population (Durbin The present study examines the idea that reproductive et al. 2003). success in E. glacialis is linked to food availability and Food limitation due to low availability of phytoplank- evaluates satellite-derived sea-surface chlorophyll ton has also been hypothesized as a source of Calanus (chl) concentration as an environmental proxy for the finmarchicus mortality (Campbell et al. 2001), specifi- nutritional potential of right-whale feeding grounds. cally in the naupliar stages that are more susceptible to The relationship between nutrition and reproduction starvation. Starvation in these early stages may ulti- in female mammals has been well documented for a mately limit the availability of high-quality right-whale number of species. In terrestrial mammals, body fat food (Stage 4 and 5 copepodites) later in the season. impacts fertility by supplying the necessary energy for Thus, those phytoplankton bloom conditions that favor reproduction and for sexual maturation (Frisch 1984, the accumulation of the right whale’s food source Thomas 1990). In the North Atlantic fin whale Bal- (C. finmarchicus) may in fact be directly beneficial for aenoptera physalus, Lockyer (1986) found that an the whales, particularly for females, who must meet apparent body-fat enhancement paralleled an increase the demands of pregnancy and nursing, through either in food supply and appeared to be associated with higher food intake or larger metabolic reserves. increased fecundity. For Eubalaena glacialis, blubber Data on the variability in phytoplankton biomass thickness can be measured acoustically in the field provide a means to explore the hypothesis that food (Moore et al. 2001) and poor body condition, as indi- limitation of Calanus finmarchicus impacts the growth cated by low blubber thickness, has been correlated and survival of Eubalaena glacialis. Surface concentra- with poor reproductive success. Angell (2005) found tion of chl, the dominant photosynthetic pigment in that blubber thickness in reproductively active female phytoplankton, is widely used as an index of phyto- right whales decreased during lactation, increased plankton abundance and biomass. Because the right after weaning, and was thickest several months prior whale’s zooplankton food source, C. finmarchicus, is to the start of pregnancy. This result emphasizes that only 1 trophic level above phytoplankton, the distribu- females utilize their blubber reserves for energetic tion and concentration of phytoplankton may reflect support during reproduction, and argues for the impor- potential C. finmarchicus production over scales re- tance of evaluating nutritional factors in regulating the lated to life-cycle progression and advective processes reproductive success of this species. in the Gulf of Maine (~10s to 100s of kilometer spatial A thorough evaluation of right-whale population scales and weekly to monthly time scales). Food avail- dynamics requires a better understanding of the spa- able during and before the right whale’s gestation tial and temporal variability of the whales’ food supply. period may be a critical factor regulating their repro- Numerous studies (Murison & Gaskin 1989, Mayo & ductive success, but this is difficult to assess directly Marx 1990, Kenney & Wishner 1995, Wishner et al. with conventional sampling approaches because of the 1995, Beardsley et al. 1996, Woodley & Gaskin 1996, large spatial ranges and extended periods of time Baumgartner & Mate 2003) conducted in each of the involved in whale feeding behavior. Remote sensing major feeding habitats in the western North Atlantic approaches for examining the spatio-temporal distrib- have demonstrated that a single species of zooplank- ution of phytoplankton biomass provide a practical ton, the 2 to 3 mm long calanoid copepod Calanus fin- means to explore a possible link between reproductive marchicus, particularly late-stage copepodites, is the success in whales (calf production) and environmental primary prey. conditions that may regulate their food supply. Recent studies have demonstrated that the abun- Remotely sensed ocean-color data provide unique dance and distribution of phytoplankton can directly synoptic views of marine phytoplankton biomass over affect egg production rates (Campbell et al. 2001) and long periods of time and large horizontal expanses of development (Crain & Miller 2001) of Calanus finmar- the surface ocean, far beyond what can be assessed chicus. For example, Campbell et al. (2001) showed from shipboard measurements. Data acquired by the that in areas of higher chl concentration, early cope- Sea-viewing Wide Field-of-view Sensor (SeaWiFS) podite stages were larger and in better condition than since its launch in 1997 can be used to quantify the copepodites found in nearby areas with lower chl con- spatial and temporal variability of phytoplankton chl (a centration. Phytoplankton blooms can have a direct proxy for biomass) (e.g. O’Reilly et al. 1998). Chl esti- effect on the late copepodite stages as well, especially mated from ocean color is an imperfect proxy for total if they persist for some time (Durbin et al. 2003). For phytoplankon biomass for reasons ranging from vari- Hlista et al.: Right whales and chlorophyll 291

ability in the ratio of phytoplankton carbon to chl to the patterns in chl variability from SeaWiFS show any presence of subsurface chl layers (~10 to 50 m depth) relationship to those of right-whale distributions and that can be below the penetration depth for satellite fecundity. observations (Gordon & Clark 1980) and may at times be important for regulating copepod distributions (e.g. Townsend et al. 1984). Despite these challenges, the MATERIALS AND METHODS advantages of mesoscale spatial and temporal coverage provided by ocean-color satellite sensors can be We examined satellite-derived chl in the major very important for exploring links between higher ‘high-use’ feeding habitats of Cape Cod Bay, Great trophic levels and primary producers. For example, South Channel, Bay of Fundy, and Roseway Basin and Platt et al. (2003) used SeaWiFS observations off Nova compared patterns to those in right-whale distributions Scotia to provide compelling evidence that larval fish and annual calving rates. Our approach uses satellite- survival is linked to the timing of the spring bloom, an derived ocean-color data for the Northwest Atlantic, important piece of the long-standing match/mismatch the extensive right-whale sightings database for the hypothesis dealing with food availability at critical region, and the photo-identification database of times after spawning (see Cushing 1975, 1990). These observed numbers of calves produced each year from kinds of hypotheses cannot be tested adequately with 1998 to 2007. conventional sampling approaches. The close trophic Ocean-color processing. The SeaWiFS archive connection between right whales and phytoplankton (August 1997 to December 2006) was obtained from and the advantages of satellite data have motivated us NASA/Goddard Space Flight Center as daily Level 1A to explore the relationship between remotely sensed MLAC (Merged Local Area Coverage) and processed phytoplankton pigments, whale sightings, and calf to high-resolution Level 2 products (chl concentration production in the Northwest Atlantic. and normalized water-leaving radiance). During Level Previous examinations of right-whale distributions 2 processing, atmospheric corrections and bio-optical and migratory patterns (Winn et al. 1986, Gaskin 1991, algorithms, OC4v4 (O’Reilly et al. 1998, 2000), were Kenney et al. 2001, Kraus et al. 2001) have shown that applied to the sensor data as implemented in SeaDAS this species typically spends the period between Janu- (SeaWiFS Data Analysis System) version 4.9 (Fu et al. ary and October feeding in the Northwest Atlantic. 1998), with flags optimized for the study area (masked Four primary feeding grounds (see Fig. 1) appear to be pixels according to SeaDAS defaults, plus flags for consistently utilized by right whales on a seasonal high solar zenith and for low normalized water-leaving basis: (1) Cape Cod Bay, (2) Great South Channel, radiance at 555 nm). On occasion, the satellite orbit (3) the southern portion of the Bay of Fundy, and (4) the configuration produced 2 scenes d–1, one of which was Roseway Basin portion of the shelf. The areal extent invariably at a wide scan angle resulting in distorted and the multi-year, multi-seasonal nature of any prob- pixels. In these cases, the satellite orbits that produced lem involving whale feeding effects emphasizes that images nearest nadir were kept in the data set and the satellite observations are the only way currently avail- distorted images were discarded, because the chl con- able to characterize phytoplankton variability on ap- centrations and distributions were unreliable. This propriate scales. processing resulted in daily scenes, spanning the Previous studies have documented strong seasonal- entire period of 1997 to 2006, with gaps in time and ity and interannual variability in chl distributions in the space principally due to cloud cover. SeaWiFS algo- Gulf of Maine and surrounding areas (Yoder et al. rithms are optimized for clear-water (Case 1) regions 2002, Thomas et al. 2003). Furthermore, previous and chl retrievals have been validated to within 30 to efforts to evaluate SeaWiFS retrievals support their use 35% over the concentration range 0.01 to 64 mg m–3 for for examining variability in this region. As emphasized open-ocean and coastal environments (Hooker & in analyses by Gregg & Casey (2004), in situ chl obser- McClain 2000, Gregg & Casey 2004, Lavender et al. vations for the Gulf of Maine region comprise a large 2004). In our analyses, we discarded retrieved chl fraction (~40%) of the available global data set for values outside of this validation range. evaluation of SeaWiFS retrievals, and retrieval perfor- Individual daily scenes were mapped to the North- mance for this region is similar to the rest of the globe East Coast projection (NEC) (see Fig. 1) at the highest (~32% RMS log error). Even in challenging sites such resolution (1 km pixel–1). The NEC projection is a stan- as semi-enclosed embayments of the Gulf of Maine, dard used by the University of Rhode Island and standard SeaWiFS retrievals are correlated with in situ National Oceanic and Atmospheric Administration values, though with some bias towards overestimates (URI-NOAA) Remote Sensing Group. The resulting from the satellite (Hyde et al. 2007). Our goal was to daily images were then used to produce weekly (8 d) build on this foundation to evaluate whether observed and monthly composites, by taking the geometric 292 Mar Ecol Prog Ser 394: 289–302, 2009

mean of the non-masked available chl concentrations SPUE values were ordered from highest to lowest. for the entire region encompassing the Eubalaena Finally, smaller, rectangular high-use areas were glacialis feeding grounds in the Northwest Atlantic. defined without compromising the sightings data, to Since the distribution of chl in shelf and slope waters include grid cells that encompassed most of the high- approximately follows a log-normal distribution density SPUE and a large portion of the total SPUE (Campbell 1995), the logarithms of chl concentrations (see Fig. 1). were taken prior to calculating the composites. Calving observations. The NARWC photo-identifi- Right-whale sightings distribution. The boundaries cation catalog, curated by the New England Aquarium of 4 ‘high-use’ areas within the feeding grounds were (NEA), contains >41 000 sighting records of individual determined with data obtained from the North Atlantic right whales as a result of intensive photographic iden- Right Whale Consortium (NARWC) sightings data- tification efforts since 1980 (Hamilton & Martin 1999). base, maintained at URI (Kenney 2001). The full data- Because females usually give birth in the calving base of sightings records extends back to the 18th cen- ground along the coast of Georgia and southeastern tury and includes >30 000 right-whale observations Florida, the photographic data resulting from intensive from aerial and shipboard surveys and opportunistic survey efforts in that area (~90 aerial survey days yr–1) sources. The NARWC data set stored and processed at and in the northern feeding grounds where mother- URI is regularly updated, but only observations from calf pairs are also sighted captures most, if not all, the available period of overlap with the available calves in the population (Kraus et al. 2001). NEA pro- SeaWiFS remote sensing imagery (1997 to 2005) were vided summary data of annual calf count, calculated analyzed for the present study. from the total number of mothers identified with calves Because of potential bias caused by uneven distribu- for the years 1998 to 2007. tion of survey effort, we used the sightings per unit- Time series. Composite satellite data were used to effort (SPUE) algorithm developed at URI (CETAP calculate time series of average chl concentration in 1982, Kenney & Winn 1986, Shoop & Kenney 1992) to each high-use area defined from the SPUE analysis. provide a quantitative assessment of right-whale dis- Data extracted from daily imagery with ≥15% valid tributions. For the present study, all aerial and ship- pixels in each high-use area were used to create board survey data, from the Gulf of Maine and sur- weekly and monthly time series over the entire period rounding waters (north of 39° N and west of 63° W), of available SeaWiFS data. Using this criterion, daily were extracted from the NARWC data set to quantify imagery for each of the high-use areas was available sampling effort and derive right-whale SPUE values. 80 to 100% of the time. On average, these weekly and The entire study region was partitioned spatially into a monthly composites provided a spatial coverage of 86 grid of cells measuring 3’ of latitude (5.6 km) by 3’ of and 96%, respectively. By taking the mean of each longitude (3.9 to 4.3 km) and temporally by month. month over the period 1997 to 2006, it was possible to Survey effort was first quantified as the length of produce mean annual cycles and explore differences track sampled under acceptable survey conditions. in seasonal patterns between the areas of interest. Only track line segments completed with clear visibil- Since we wanted to explore general relationships ity of ≥2 nautical miles (3.7 km), sea state ≤ Beaufort 3, between chl time series and right-whale sightings, aircraft altitude <1200 feet (366 m), and ≥1 observers monthly resolved SPUE data for each high-use area on watch were included as acceptable effort. Similarly, were also analyzed to produce mean annual cycles of only right whales sighted during acceptable effort whale distributions. These annual cycles were calcu- were included and summed within each cell and lated in the same way as for the SeaWiFS data, i.e. by month. Finally, the number of right whales sighted was computing the mean for each month given all the divided by effort to generate the SPUE index, in units observations over the period from 1997 to 2006. of whales sighted per 1000 km of valid effort. Annual weighted chl (nutritional index). Initial find- These data were used to specify boundaries that ings relating seasonal patterns in mean annual chl and reflect the most widely used feeding grounds around SPUE across the high-use areas were used to calculate Cape Cod Bay, Great South Channel, Bay of Fundy, a simple nutritional index to capture annual variations and Roseway Basin. Initially, areas substantially larger in chl that might be relevant to the right-whale popula- than the federally defined right-whale critical habitats tion. With the goal of generating an index that reflects (USA) or right-whale conservation areas (Canada) aspects of average potential food condition in each of were created around each of the 4 feeding areas, such the high-use areas during the season of peak sightings, that all of the 1997 to 2005 sightings were included. we calculated a simple 2-stage average chl concentra- Effort and sightings across all years and months were tion. First, for a given year, chl was averaged over all of summed within each grid cell, and a single overall the pixels within each high-use area and over the SPUE value per cell was computed. The non-zero period of relatively high concentration specific to each Hlista et al.: Right whales and chlorophyll 293

area (see details below). Then, the resulting values reproduction (Angell 2005) and ≥3 yr are typically (1 for each of the 3 high-use areas; Roseway Basin was required between successive calving: 1 yr for lactation, not included in the average due to sparsity of SPUE a resting year to accumulate fat stores, and a gestation data) were averaged to obtain a single weighted-aver- year (Knowlton et al. 1994). Thus, we evaluated 3 ca- age chl concentration for the year. This value is ses: (1) the nutritional index in the year prior to calving, weighted such that only the overlap between ‘bloom’ (i.e. gestation), (2) the nutritional index 2 yr prior to period and peak sightings is considered in a way that is calving (i.e. resting), and (3) the mean of the nutritional specific to each high-use area defined by the SPUE index in both years prior to calving. In each case we analysis. For example, in Cape Cod Bay, only early used reduced major-axis Model II regression (Sokal & months of the year contribute to the average because Rohlf 1981) to summarize the relationships between chl concentration and whale sightings tend to be high- chl and calf production. est in that period. A more complex average could reflect seasonal variability in prey composition or weightings that reflect habitat value (i.e. if one area RESULTS was known to be a more important feeding ground than others, its contribution could be weighted more Right-whale sightings distributions heavily). Given the lack of information to support the inclusion of these additional factors, however, we used The 4 high-use areas as defined by the analysis of a simple equal weight for each area. the SPUE data delineate regions that encompass the To explore the relationship between the annual most frequent right-whale sighting locales: in the weighted chl and right-whale reproductive success vicinity of Cape Cod Bay, the Great South Channel, the (number of calves observed), we considered the effect Bay of Fundy, and Roseway Basin (Fig. 1). For the of several time lags. Variations in food availability can period 1997 to 2005, these defined areas exclude only be integrated over multiple years because females uti- 5 to 20% of the cumulative sightings (i.e. relatively lize their blubber reserves for energetic support during rare sightings in surrounding areas). While the high-

Fig. 1. High-use areas (black outline) defined by analysis of the sightings per unit effort (SPUE) data for North Atlantic right whales encompass the pre-existing critical habitat or conservation zones (red outline). Background image shows July monthly mean composite chlorophyll (chl) distribution from analysis of SeaWiFS observations. CCB: Cape Cod Bay, GSC: Great South Channel, BOF: Bay of Fundy, RB: Roseway Basin 294 Mar Ecol Prog Ser 394: 289–302, 2009

use areas we defined extend beyond the bounds cur- (Fig. 1, Table 1), they are consistent with previous rently designated as critical habitat (USA) or conserva- analyses of the sightings database (Winn et al. 1986, tion zone (Canada) for the North Atlantic right whale Kenney et al. 2001). Examination of average right- whale SPUE for each year from Table 1. Properties of the 4 high-use areas defined by analysis of sightings per unit 1997 to 2005 shows that, for each effort (SPUE) data. Latitudes and longitudes for northern, southern, western, and area, the arrival and departure of eastern edges specify the areas whales vary considerably from year to year (Fig. 2). From this analysis, it High-use North South West East No. ofAreal is also clear that arrivals and depar- area pixels extent tures are not always well resolved (km2) on the basis of the first and last sur- Cape Cod Bay 42.25° N 41.70° N 70.55° W 70.00° W 1900 41 veys of the season. It is important, Great South Channel 42.50° N 41.00° N 69.70° W 68.15° W 14098 113 however, to note that all years show Bay of Fundy 44.95° N 44.25° N 66.70° W 66.20° W 2170 44 a consistent pattern of transition Roseway Basin 43.15° N 42.70° N 65.90° W 65.00° W 2356 46 from Cape Cod Bay (January to May) to the Great South Channel (March to June) to the Bay of Fundy (June to October) as the seasons progress from winter to summer (Fig. 3). This view is consistent with previous analyses of right-whale sightings distributions (Winn et al. 1986, Kenney & Wishner 1995, Ken- ney et al. 2001). Cape Cod Bay. SPUE data for the period from 1997 to 2005 show few whales present as early as Novem- ber or December in Cape Cod Bay; these may represent early arrivals for the new feeding season or late departures (on their way south- ward) from the previous year (Allen 1916). An exception is 1998 when the mean SPUE was very high (83 whales per 1000 km in December), possibly associated with a shorter- than-usual survey that happened to coincide with high whale density. More typically, the SPUE data suggest the abundance of whales begins to increase in January and reaches a maximum between Feb- ruary and April. By May, right whales have consistently departed Cape Cod Bay (Figs. 2 & 3). Great South Channel. Interan- nual variability is notable in arrival and departure of whales from Great South Channel (Figs. 2 & 3). None- theless, some generalities emerge. There are consistently few whales Fig. 2. Average right-whale sightings per unit effort (SPUE) in each high-use area: sighted in March and the observa- (a) Cape Cod Bay, (b) Great South Channel, (c) Bay of Fundy, and (d) Roseway tions suggest that most right whales Basin, for each year from 1997 to 2005 begin arriving during April, peak in Hlista et al.: Right whales and chlorophyll 295

Roseway Basin. Roseway Basin, an area previously identified as an important habitat for right whales, was charac- terized by relatively high SPUE levels. However, during the period of the present study, the distribution of survey effort was very uneven with poor temporal coverage. Unfortunately, the sparse SPUE data in Roseway Basin precludes a detailed examination of this area for the quantitative assessment of right-whale habitat characteristics (Fig. 2). Thus, this area was not included in the development of the cumulative right-whale nutritional index.

SeaWiFS chl variability

We used time series of chl extracted from weekly and monthly (Fig. 4) composite SeaWiFS observations to examine the high concentration of phytoplankton (or bloom period) for each year from 1998 to 2006 in the 4 high-use areas that were defined by analysis of SPUE distributions (black outlines in Fig. 1). Although the exact timing and duration of blooms vary within each region, there is an overall seasonal progression, with highest chl concentration shifting from Cape Cod Bay to the Great South Channel to the Bay of Fundy as the seasons progress from winter to summer (Fig. 5). Cape Cod Bay. Cape Cod Bay is char- acterized by presence of both spring and Fig. 3. Average right-whale sightings per unit effort (SPUE) in each of the high- fall blooms, although their prominence use areas, grouped by year for (a) 1997, (b) 1998, (c) 1999, (d) 2000, (e) 2001, (f) 2002, (g) 2003, (h) 2004, and (i) 2005. All years show a consistent pattern of varies from year to year (Fig. 4). The tim- transition from Cape Cod Bay (—•—) to Great South Channel (—•—) to Bay of ing of onset and decline of the blooms is Fundy (—•—) as the seasons progress from winter through summer. Mean also variable from year to year. Typically, SPUE data for Cape Cod Bay and the Great South Channel are shown with a lowest chl concentrations occur in sum- 4-fold different scale (no. per 4000 km) than that for the Bay of Fundy (no. per mer between the bloom periods. Of all 1000 km) to emphasize transitions the areas, Cape Cod Bay has the highest chl concentrations. May and June, and depart by July. The year 2005 was Great South Channel. The dominant feature in the an anomalous year in that right whales arrived earlier Great South Channel is a spring bloom (Fig. 4). The and abundance was exceptionally high in April. amplitude varies from year to year, but the highest chl Bay of Fundy. Because of the distribution of survey concentrations always occur in April. Very prominent effort (almost none in May or November), arrival to blooms occurred in 1999, 2000, and 2004. and departure from the Bay of Fundy are not well Bay of Fundy. A distinct feature in the Bay of Fundy resolved in this data set (Figs. 2 & 3). The SPUE data do is an extended summer bloom (Fig. 4). These blooms show, however, that right whales are most abundant last ≥6 mo, often with 2 distinguishable maxima, usu- between July and September for each year from 1997 ally separated by 3 to 4 mo. In this area, the lowest chl to 2005. Interestingly, SPUE values in the Bay of Fundy concentrations occur during the period spanning late are the highest of the 4 high-use habitat areas. fall to early spring. 296 Mar Ecol Prog Ser 394: 289–302, 2009

peak chl concentrations. Within the broadly defined feeding season (late winter through early fall), highest concentrations are observed throughout the winter and early spring in Cape Cod Bay, from March to May in the Great South Channel, and throughout the summer in the Bay of Fundy (Fig. 5, left panels).

Annual weighted chl index and calf count

While patterns of whale distribution may be related to chl distributions on seasonal time scales, we expect annual or longer time scales to influence reproductive success (number of calves observed). This led us to consider an annual weighted chl index. We used the details of the mean annual cycles to select the time ranges contributing to the annual weighted chl concentration. The bloom periods, capturing highest chl concentrations, were averaged within Cape Cod Bay (January to April), the Great South Channel (March to May), and the Bay of Fundy (May to October), such that the relevant monthly period is coincident with peak right-whale sightings in each of the high-use areas. The resulting annual weighted chl concentration (which represents 1 realization of a nutri- Fig. 4. For each high-use area: (a) Cape Cod Bay, (b) Great South tional index) shows interannual variability Channel, (c) Bay of Fundy, and (d) Roseway Basin, 9 yr time series of over the period 1997 to 2006 with an intrigu- average chlorophyll (chl) concentration, extracted from monthly compo- ing relationship to the number of right whale site SeaWiFS observations calves observed (Fig. 6). There is a significant positive correlation between the number of Roseway Basin. During the present study period, calves produced and weighted chl averaged over 2 Roseway Basin typically had the lowest chl of all the prior years (Fig. 7c, r = 0.71, p = 0.049). There are also areas, with concentrations especially low during win- positive but not significant (at p = 0.05) correlations for ter (Fig. 4). In most years, Roseway Basin has spring weighted chl with a 1 yr lag (Fig. 7a, r = 0.39, p = 0.306) and fall blooms, though 2000 was an anomalous year in and 2 yr lag (Fig. 7b, r = 0.66, p = 0.075). that no spring bloom occurred. The 2 yr average of annual weighted chl concentration (encompassing the whale’s resting year and gestation year) may have predictive capability. For example, Mean seasonal cycle comparison the highest calf numbers, in 2001 and 2005, were observed 1 yr after the 2 yr average of highest Comparison of the monthly-resolved mean annual weighted chl values (1999 to 2000 and 2003 to 2004). cycle for SeaWiFS-derived chl and SPUE in each of the For the entire period of study (all points included), the high-use areas shows notable patterns. The results of 2 yr weighted chl explains 50% of the variance in the the SPUE analysis, for the period from 1997 to 2005, number of calves produced the next year (Fig. 7c). emphasize that highest density right-whale sightings progress from Cape Cod Bay to the Great South Chan- nel to the Bay of Fundy (Fig. 5, right panels). The sight- DISCUSSION ings are too sparse in Roseway Basin to constrain a mean annual cycle for these years. This pattern of tran- The results of the right-whale SPUE analysis high- sition is consistent with the seasonal progression of light seasonal time periods of dense whale sightings in Hlista et al.: Right whales and chlorophyll 297

with the Middle Atlantic Bight than the Gulf of Maine (Yoder et al. 2002). Winn et al. (1986) were the first to pro- pose a 6-phase model of the right whale’s annual migratory cycle. The model des- cribes the general population movements of the right whales. During the winter, most reproductively active females give birth along the coasts of Georgia and northeastern Florida. Non-calving females and males, on the other hand, are rarely seen in the calving grounds and their locations during the early winter remain largely un- known (Kenney et al. 2001). As the seasons progress, the reproducing females migrate northward and join with the rest of the population in the late winter or early spring, re- maining in the feeding habitats through summer and fall. Although variability is evident, our analysis of recent right-whale SPUE distribution is consistent with the movements and patterns initially set forth in Winn et al.’s (1986) framework. To sustain their large metabolic demands, right whales migrate to habitat areas containing their food source and for- age on dense prey patches. Winn et al. (1986) hypothesized that the seasonal progression of the right whale’s migratory cycle coincides with high-density concentrations of their principal prey, Calanus fin- Fig. 5. Comparison of monthly-resolved mean cycle for SeaWiFS-derived marchicus. Numerous studies conducted in chlorophyll (chl) and for sightings per unit effort (SPUE) in each of the high- each of the major feeding habitats have use areas: (a,b) Cape Cod Bay (CCB), (c,d) Great South Channel (GSC), (e,f) confirmed that right whale’s feeding activ- Bay of Fundy (BOF), and (g) Roseway Basin (RB). Highest density right- ity is directed towards locating and exploit- whale sightings progress from CCB to GSC to BOF and are consistent with ing dense patches of C. finmarchicus the seasonal progression of highest chl concentrations (Baumgartner & Mate 2003, Baumgartner et al. 2003). Though exciting progress is 3 high-use areas in the Northwest Atlantic: Cape Cod being made, the oceanographic conditions that pro- Bay, the Great South Channel, and the Bay of Fundy. mote dense patches of this prey over time and space For the period 1997 to 2005, whale sightings are most remain poorly understood, as do the migration and for- abundant in these 3 regions during characteristic aging strategies the right whales use to locate and months and show seasonal progression from Cape Cod exploit these food resources (Kenney et al. 2001). Nev- Bay to the Great South Channel to the Bay of Fundy. ertheless, we do know that, during late winter and This seasonal progression is consistent with the transi- early spring, right whales are found feeding in coastal tion of peak satellite-derived chl concentrations in waters of Cape Cod Bay when their zooplankton food each of the high-use areas that were defined by the source is at a maximum, and whales typically leave the SPUE analysis (Fig. 6). The range of patterns in chl sea- area, presumably in search of a better food supply, as sonality evident in these areas is similar to previous their prey sources start to decrease (Mayo & Marx reports by Thomas et al. (2003) for a different selection 1990). Likewise, right whales feed in the Great South of geographic locations within and around the Gulf of Channel during late spring (Kenney & Wishner 1995, Maine. When compared to Thomas et al.’s (2003) loca- Beardsley et al. 1996) and the Bay of Fundy during the tions, our results highlight Cape Cod Bay as an summer months (Murison & Gaskin 1989, Woodley & extreme case of relatively high wintertime concentra- Gaskin 1996, Baumgartner & Mate 2003) when high- tions, a seasonal pattern generally more consistent density concentrations of C. finmarchicus occur. 298 Mar Ecol Prog Ser 394: 289–302, 2009

tion. Copepods in the Gulf of Maine are well known to form deep layers, especially late in the season when right whales have also been documented to exploit those layers (Baumgartner & Mate 2003, Baumgart- ner et al. 2007). The intensity and timing of near-surface blooms may influence deep layers but through a complex set of factors not explicit in our analysis. Furthermore, surface chl concentrations and copepod aggregations are both influenced by some of the same oceanographic conditions, such as level of stratification and presence of fronts that vary with season and year. It may be that relationships between whales and chl patterns emerge because of this kind of covariance. When we averaged annual weighted chl concentration over the 2 yr period prior to calving, encompassing the resting year Fig. 6. Time series of number of calves observed (solid line) and annual and the gestational year for the reproduc- weighted SeaWiFS chl (dashed line). The 2 yr with highest calf numbers tive females, half of the variation in breed- are observed 1 to 2 yr after the highest weighted chl values. Chl values are ing success can be explained (Fig. 7c). weighted such that bloom periods, capturing highest chl concentrations, were averaged within Cape Cod Bay (January to April), the Great South There is also a suggestive trend with a sin- Channel (March to May), and the Bay of Fundy (May to October) to ensure gle 1 yr lag, but with a strong outlier year that the relevant monthly period overlaps with peak whale sightings (calves in 2000) for that case (Fig. 7a). Con- in each of the high-use areas defined by the sightings per unit effort sidering potential food availability over the (SPUE) analysis 2 yr span may shed some light on the patterns that are not well explained by a sin- The right whale’s zooplankton food source, Calanus gle-year lag. For instance, the lowest number of calves finmarchicus, is a direct consumer of phytoplankton. (1 calf in 2000) was observed 2 yr after the lowest Therefore, the timing of phytoplankton abundance annual weighted chl value (i.e. chl observations in and biomass development in each of the right-whale 1998, during the whale’s resting period). Even though feeding habitats could have important consequences, high chl conditions returned in 1999, the time-lagged especially since C. finmarchicus shows signs of food response (to successful reproduction) may have been limitation in areas of low surface-chl concentration hampered by extremely low food availability during (Campbell et al. 2001). We found that the temporal pat- the whale’s resting period, possibly delaying concep- tern of right-whale movement between feeding areas tion. Periods of extreme food limitation might affect a is consistent with changes in satellite-derived chl con- female’s response to previous successful reproduction centrations in those areas, which may reflect an impact by prolonging the crucial time needed to recover. It is through C. finmarchicus populations. As the seasons likely that a threshold food requirement and associ- progress from winter to summer, the highest chl con- ated body condition during the resting year must be centrations are observed throughout winter and early met before successful conception and early pregnancy spring in Cape Cod Bay, from March to May in the can occur. If food conditions in a given year are below Great South Channel, and throughout the summer in this threshold (as may be the case during 1998, with the Bay of Fundy. While we cannot quantify the effect weighted chl <2 mg m–3), then it is possible that food of this progression on C. finmarchicus distributions, it conditions during the subsequent year do not control seems likely that post-bloom conditions will impact calf production since the minimum requirements for C. finmarchicus, for instance contributing to spatial conception and early pregnancy were not met. Longer and temporal variability in development rate or the time series are needed to evaluate whether these kinds onset of diapause (e.g. Durbin et al. 2000, 2003). We of factors contribute to consistent patterns in the rela- emphasize that our analysis and interpretation is not tionship between chl and reproductive success (num- meant to imply that C. finmarchicus (and right whales ber of calves observed). in turn) consistently feed on local surface-layer phyto- When calf numbers are lagged by 1 yr (Fig. 7a), plankton that are accessible to satellite-based observa- weighted chl concentration has less predictive power Hlista et al.: Right whales and chlorophyll 299

than when a 2 yr lag is considered (Fig. 7b), with weighted chl concentration (during the mother’s resting year) explaining 43% (p = 0.075) of the variance in number of calves produced; this result may reflect that gestation itself is not especially costly, while the energetic demands for lactation are substantial. This idea is consistent with observations for another cetacean, the harbor porpoise, in which costs of lactation are 3 times greater than for pregnancy (Yasui & Gaskin 1986). Given the fact that mother right whales utilize their blubber reserves for energetic support during reproduction (Angell 2005), it is possible that they do not conceive until they have enough energy reserved to make it through lactation. Because of the challenges studying cetaceans in their natural habitats, detailed mechanisms leading to reproductive delays (i.e. failure to conceive, failure to carry to term, low survival of neonates) remain poorly constrained, but there is well- documented variability in calving interval (Kraus et al. 2007) and established links between body condition and reproduction in right whales (Moore et al. 2001, Angell 2005). Our findings provide further support for a general link with food availability. It is surprising that a single factor, such as 2 yr weighted chl, might explain so much of the variation in right-whale reproductive success. Right-whale biology and ecological interactions are well known to be more complex than this view encompasses. Nonetheless, our results support the view that surface chl concentration, especially as sampled with the resolution possible from earth-orbiting satellites, provides a practical means to integrate broadly over a range of factors directly affec- ting the right-whale population. It may not be necessary to resolve all the links or complexity involved in the relationship between primary producers and whales to identify relevant indices. The underlying premise motivating the present study is that reproductive success in the North Atlantic right whale is limited by food availability for which surface chl conditions may be a working proxy. The evidence of a relationship between 2 yr average weighted chl (i.e. a nutritional index) and the number of new right-whale calves is consistent with this idea. For the present study period, the most successful years for reproduction (i.e. 2001 and 2005) follow years with highest weighted chl values averaged over 2 yr, and unsuccessful years (i.e. 1999 and 2000) occur when weighted chl was very low during at least 1 of the 2 Fig. 7. Relationship between number of calves observed and preceding years. Years of extreme food limitation may annual weighted chlorophyll (chl) with (a) 1 yr lag in calf delay a mother’s response to successful reproduction numbers, (b) 2 yr lag in calf numbers, and (c) chl averaged by prolonging the crucial time needed to store surplus over the 2 yr prior to calving. With a 1 yr lag and a 2 yr lag, energy for the next healthy pregnancy. weighted chl explains 14% (p = 0.306) and 43% (p = 0.075) of The mechanistic links between chl and right-whale the variance in the number of calves, respectively. The 2 yr averaged chl explains 50% (p = 0.049) of the variance in the feeding success are complicated and mediated by a number of calves produced the next year variety of biological and environmental processes. 300 Mar Ecol Prog Ser 394: 289–302, 2009

They depend on many factors, including the complex also facilitate further applications of high-resolution life history and ecology of prey species such as Cala- remotely sensed observations. For instance, new sam- nus finmarchicus and behavioral and physical oceano- pling technologies in the form of passive acoustic graphic conditions that affect prey aggregation into devices (‘pop-ups’) have been employed in the field to dense patches critical for effective whale feeding (e.g. study the seasonal occurrence and distribution of right Baumgartner et al. 2007). These factors can be quanti- whales in the North Atlantic (Parks & Clark 2005). fied only in the context of detailed process studies, These passive acoustic methods have the potential to which are necessarily limited in spatial and temporal augment visual observations from aerial and shipboard resolution. The notable advantage of remote-sensing surveys, enabling a sharper definition of the arrival approaches is that they allow wide ranges of time and and departure times for each habitat. The integration space to be assessed. Future efforts that exploit a of whale distribution data, from both visual and combination of these approaches may provide new acoustic measures, with concurrent satellite-derived insights. chl observations, will provide a better understanding While the present study is based on a relatively sim- of the migration patterns and habitats of right whales, ple analysis of the first 9 yr of ocean-color data ob- ultimately assisting in the management of this highly tained from the SeaWiFS mission, these data provide a endangered species. unique synoptic quantification of seasonal variability in chl concentration within the right-whale feeding Acknowledgements. We appreciate the commitment of S. D. Kraus, director of the right-whale research program at grounds in the Northwest Atlantic. A longer time series the New England Aquarium, and the wonderful research is critical to examine the generality of predictive rela- groups who are dedicated to monitoring right-whale popula- tionships between weighted chl concentration and calf tions, including calf production. Access to the whale sightings production. Data from the next generation of global and survey database at University of Rhode Island and the calving summary data at the New England Aquarium were ocean-color missions (e.g. moderate-resolution imag- provided through the North Atlantic Right Whale Consortium ing spectroradiometer [MODIS], medium-resolution (www.rightwhaleweb.org). The bulk of the surveys that went imaging spectrometer [MERIS]) will provide an oppor- into the sightings per unit effort data were conducted by the tunity for further analysis. These new satellite data New England Aquarium (Boston, Massachusetts), Province- sets, which overlap with and are expected to extend town Center for Coastal Studies (Provincetown, Massachu- setts), National Marine Fisheries Service Northeast Fisheries beyond the SeaWiFS mission, could improve our Science Center (Woods Hole, Massachusetts), and East Coast understanding of right-whale reproduction and its Ecosystems Research Organization (Pierrefonds, Quebec). environmental regulation. Longer time series will also This research was supported by grants from the Northeast enable more complex uses of satellite-derived chl ob- Consortium and the Ocean Biology and Biogeochemistry Program at NASA. servations. For instance, it may be desirable to en- hance demographic models (e.g. Fujiwara & Caswell 2001) to explicitly incorporate a remotely sensed mea- LITERATURE CITED sure of food availability and to represent its impact on Allen GM (1916) The whalebone whales of New England. particular life-cycle stages (Greene et al. 2003). This Mem Boston Soc Nat Hist 8:107–322 kind of approach would allow us to move away from Angell CM (2005) Body fat condition of free-ranging right simple correlation analyses and to more carefully con- whales, Eubalaena glacialis and Eubalaena australis. PhD sider the biological and ecological implications of food dissertation, Boston University Baumgartner MF, Mate BR (2003) Summertime foraging ecol- availability on right-whale populations. ogy of North Atlantic right whales. Mar Ecol Prog Ser More complex correlations among basin-scale phe- 264:123–135 nomena associated with changes in the North Atlantic Baumgartner MF, Cole TVN, Campbell RG, Teegarden GJ, Oscillation, regional physical oceanography, and the Durbin EG (2003) Associations between North Atlantic abundance of Calanus finmarchicus within the right- right whales and their prey, Calanus finmarchicus, over diel and tidal time scales. Mar Ecol Prog Ser 264:155–166 whale feeding grounds reveal that climate-driven Baumgartner MF, Mayo CA, Kenney RD (2007) Enormous changes may account for some of the variability in carnivores, microscopic food, and a restaurant that’s hard C. finmarchicus abundance, with possible consequen- to find. In: Kraus SD, Rolland R (eds) The urban whale: ces for right-whale calving rates (Greene & Pershing North Atlantic right whales at the crossroads. Harvard University Press, Cambridge, MA, p 138–171 2004, Baumgartner et al. 2007). With a longer time Beardsley RC, Epstein AW, Chen C, Wishner KF, Macaulay series of sea-surface chl, it will be possible to extend MC, Kenney RD (1996) Spatial variability in zooplankton these analyses to include relevant aspects of phyto- abundance near feeding right whales in the Great South plankton variability and to investigate the strength of Channel. Deep-Sea Res II 43:1601–1625 bottom-up control. Campbell JW (1995) The lognormal distribution as a model for bio-optical variability in the sea. J Geophys Res 100: Emerging approaches for characterizing spatial and 13237–13254 temporal variability in right-whale distributions may Campbell RG, Runge JA, Durbin EG (2001) Evidence for food Hlista et al.: Right whales and chlorophyll 301

limitation of Calanus finmarchicus production rates on the ductivity experiment. Cont Shelf Res 15:373–384 southern flank of Georges Bank during April 1997. Deep- Kenney RD, Mayo CA, Winn HE (2001) Migration and forag- Sea Res II 48:531–549 ing strategies at varying spatial scales in western North Caswell H, Fujiwara M, Brault S (1999) Declining survival Atlantic right whales: a review of hypotheses. J Cetacean probability threatens the North Atlantic right whale. Proc Res Manag 2(Spec Issue):251–260 Natl Acad Sci USA 96:3308–3313 Knowlton AR, Kraus SD (2001) Mortality and serious injury of CETAP (Cetacean and Turtle Assessment Program) (1982) A northern right whales (Eubalaena glacialis) in the western characterization of marine mammals and turtles in the North Atlantic Ocean. J Cetacean Res Manag 2(Spec mid- and north Atlantic areas of the U.S. outer continental Issue):193–208 shelf. Report no. AA551-CT8–48. Bureau of Land Man- Knowlton AR, Kraus SD, Kenney RD (1994) Reproduction in agement, Washington, DC North Atlantic right whales, Eubalaena glacialis. Can J Crain JA, Miller CB (2001) Effects of starvation on intermolt Zool 72:1297–1305 development in Calanus finmarchicus copepodites: a com- Kraus SD, Hamilton PK, Kenney RD, Knowlton AR, Slay C parison between theoretical models and field studies. (2001) Status and trends in reproduction of the North Deep-Sea Res II 48:551–566 Atlantic right whale. J Cetacean Res Manag 2(Spec Issue): Cushing DH (1975) Marine ecology and fisheries. Cambridge 231–236 University Press, Cambridge Kraus SD, Brown MW, Caswell H, Clark CW and others Cushing DH (1990) Plankton production and year-class (2005) North Atlantic right whales in crisis. Science 309: strength in fish populations: an update of the match/mis- 561–562 match hypothesis. Adv Mar Biol 26:249–293 Kraus SD, Pace RM III, Frasier TR (2007) High investment, Durbin EG, Garrahan PR, Casas MC (2000) Abundance and low return: the strange case of reproduction in Eubalaena distribution of Calanus finmarchicus on the Georges Bank glacialis. In: Kraus SD, Rolland RM (eds) The urban whale: during 1995 and 1996. ICES J Mar Sci 57:1664–1685 North Atlantic right whales at the crossroads. Harvard Durbin EG, Campbell RG, Casas MC, Ohman MD, Niehoff B, University Press, Cambridge, MA, p 172–199 Runge J, Wagner M (2003) Interannual variation in phyto- Lavender SJ, Pinkerton MH, Froidefond JM, Morales J, Aiken plankton blooms and zooplankton productivity and abun- J, Moore GF (2004) SeaWiFS validation in European dance in the Gulf of Maine during winter. Mar Ecol Prog coastal waters using optical and bio-geochemical mea- Ser 254:81–100 surements. Int J Remote Sens 25:1481–1488 Frisch RE (1984) Body fat, puberty and fertility. Biol Rev Lockyer C (1986) Body fat condition in Northeast Atlantic fin Camb Philos Soc 59:161–188 whales, Balaenoptera physalus, and its relationship with Fu G, Baith KS, McClain CR (1998) SeaDAS: the SeaWiFS reproduction and food resource. Can J Fish Aquat Sci 43: data analysis system. Proc 4th Pacific Ocean Remote Sens- 142–147 ing Conf, Qingdao, China, July 28–31, 1998, p 73–79 Mayo CA, Marx MK (1990) Surface foraging behavior of the Fujiwara M, Caswell H (2001) Demography of the endan- North Atlantic right whale, Eubalaena glacialis, and gered right whale. Nature 414:537–541 associated zooplankton characteristics. Can J Zool 68: Gaskin DE (1991) An update on the status of the right whale, 2214–2220 Eubalaena glacialis, in Canada. Can Field Nat 105: Moore MJ, Miller CA, Morss MS, Arthur R and others (2001) 198–205 Ultrasonic measurement of blubber thickness in right Gordon HR, Clark DK (1980) Remote sensing optical proper- whales. J Cetacean Res Manag 2(Spec Issue):301–309 ties of a stratified ocean: an improved interpretation. Appl Murison LD, Gaskin DE (1989) The distribution of right Opt 20:3426–3428 whales and zooplankton in the Bay of Fundy, Canada. Can Greene CH, Pershing AJ (2004) Climate and the conservation J Zool 67:1411–1420 biology of North Atlantic right whales: the right whale at O’Reilly JE, Maritorena S, Mitchell BG, Siegel DA and others the wrong time? Front Ecol Environ 2:29–34 (1998) Ocean color chlorophyll algorithms for SeaWiFS. Greene CH, Pershing AJ, Kenney RD, Jossi JW (2003) Impact J Geophys Res 103:24937–24954 of climate variability on the recovery of endangered North O’Reilly JE, Maritorena S, Siegel DA, O’Brien M and others Atlantic right whales. Oceanography 16:98–103 (2000) Ocean color chlorophyll a algorithms for SeaWiFS, Gregg WW, Casey NW (2004) Global and regional evaluation OC2, and OC4, Version 4. In: Hooker SB, Firestone ER of the SeaWiFS chlorophyll data set. Remote Sens Environ (eds) SeaWiFS postlaunch technical report series, Vol 11. 93:463–479 SeaWiFS postlaunch calibration and validation analyses, Hamilton PK, Martin SM (1999) A catalog of identified right Part 3. NASA, Greenbelt, MD, p 9–23 whales from the North Atlantic: 1935–1997. New England Parks SE, Clark CW (2005) Variation in acoustic activity of Aquarium, Boston, MA North Atlantic right whales in three critical habitat areas Hooker SB, McClain CR (2000) The calibration and validation in 2004. J Acoust Soc Am 117:2469 of SeaWiFS data. Prog Oceanogr 45:427–465 Platt T, Fuentes-Yaco C, Frank KT (2003) Spring algal bloom Hyde KJW, O’Reilly JE, Oviatt CA (2007) Validation of Sea- and larval fish survival. Nature 423:398–399 WiFS chlorophyll a in Massachusetts Bay. Cont Shelf Res Reeves RR, Rolland R, Clapham PJ (2001) Causes of repro- 27:1677–1691 ductive failure in North Atlantic right whales: new IWC (International Whaling Commission) (2001) Report of the avenues of research. Report of a workshop held 26–28 workshop on status and trends of western North Atlantic April 2000, Falmouth, MA. Ref. doc. 01–16. Northeast right whales. J Cetacean Res Manag 2(Spec Issue):61–87 Fisheries Science Center, Woods Hole, MA Kenney RD (2001) The North Atlantic Right Whale Consor- Shoop CR, Kenney RD (1992) Seasonal distributions and tium databases. Maritimes 43:3–5 abundances of sea turtles off the northeastern United Kenney RD, Winn HE (1986) Cetacean high-use habitats of States. Herpetol Monogr 6:43–67 the northeast U.S. continental shelf/slope areas. Fish Bull Sokal RR, Rohlf J (1981) Biometry: the principles and practice (Wash DC) 84:345–357 of statistics in biological research. WH Freeman, San Fran- Kenney RD, Wishner KF (1995) The south channel ocean pro- cisco, CA 302 Mar Ecol Prog Ser 394: 289–302, 2009

Thomas VG (1990) Control of reproduction in animal species Wishner KF, Schoenherr JR, Beardsley R, Chen C (1995) with high and low body fat reserves. In: Frisch RE (ed) Abundance, distribution and population structure of the Adipose tissue and reproduction, Vol 14. Karger, Basel, copepod Calanus finmarchicus in a springtime right whale p27–41 feeding area in the southwestern Gulf of Maine. Cont Thomas AC, Townsend DW, Weatherbee R (2003) Satellite- Shelf Res 15:475–507 measured phytoplankton variability in the Gulf of Maine. Woodley TH, Gaskin DE (1996) Environmental characteristics Cont Shelf Res 23:971–989 of North Atlantic right and fin whale habitat in the lower Townsend DW, Cucci CL, Berman T (1984) Subsurface Bay of Fundy, Canada. Can J Zool 74:75–84 chlorophyll maxima and vertical distribution of zooplank- Yasui WY, Gaskin DE (1986) Energy budget of a small ton in the Gulf of Maine. J Plankton Res 6:793–802 cetacean. Ophelia 25:183–197 Winn HE, Price CA, Sorensen PW (1986) The distributional Yoder JA, Schollaert SE, O’Reilly JE (2002) Climatological biology of the right whale (Eubalaena glacialis) in the phytoplankton chlorophyll and sea-surface temperature western North Atlantic. Rep Int Whal Comm Spec Issue patterns in continental shelf and slope waters off the 10:129–138 Northeast U.S. coast. Limnol Oceanogr 47:672–682

Editorial responsibility: Matthias Seaman, Submitted: October 25, 2007; Accepted: August 11, 2009 Oldendorf/Luhe, Germany Proofs received from author(s): November 8, 2009